This application is based upon and claims the benefit of priority from Japanese patent application No. 2006-189056, filed on Jul. 10, 2006, the disclosure of which is incorporated herein in its entirety by reference.
The present invention relates to a blade server which quickens the start of application services, and a method therefor.
In a system comprised of a plurality of servers, each server individually manages failure-prone portions such as a hard disk, power supply, and fan, and the reliability of the overall system cannot be increased.
To solve this, blade servers in which failure-prone portions are minimized have been developed.
The blade server comprises a common unit which gathers failure-prone portions, and a plurality of blades each having elements such as a memory and microprocessor necessary for a computer on one board.
Of these blades, the common unit, especially a management blade which manages a hard disk manages data (to be referred to as “service data” hereinafter) necessary for the services of applications activated in the remaining blades (to be referred to as “managed blades” hereinafter). The managed blade requests necessary service data of the management blade, and starts application services on the basis of the acquired service data (see, e.g., Japanese Patent Laid-Open Nos. 2005-293430, 2005-339528, and 2006-085209).
If managed blades simultaneously request service data, access concentrates on the management blade to cause communication congestion and congestion to the hard disk of the management blade.
There has been proposed a technique of preventing the congestion by expanding service data in advance in a memory managed by each managed blade and accessing the memory at the start of application services. This technique can increase the service data read speed, and can also quicken the start of application services.
According to this technique, expansion of all service data managed by the management blade into the memory of each managed blade consumes a large memory capacity. Hence, the memory consumption of each managed blade is minimized by expanding only service data necessary to start application services in each managed blade.
However, service data expanded in the memory of each managed blade change every time an application is added/deleted to/from a managed blade or the contents of service data change. The administrator of the blade server must set again in advance service data to be expanded in the memory of each managed blade. The setting is cumbersome to the administrator.